Solar charge controller recommendations for LiFePO4

[thegoogler]

Hi folks,

New-ish guy here. I've been doing SLA and solar on my trailer for year, but I've been reading a lot of the forum posts for the last couple months and started posting recently. I finished building my LiFePO4 battery bank over the weekend, and in an effort to (over)optimize my setup I wanted to pose a couple questions to those who are smarter and far more experienced with LFP than me.

My 12V setup is 4x 271Ah CATL cells (LiitoKala "280Ah") with a pseudo-overkill solar BMS (only because overkill was out of stock). I have a Renogy Rover Elite charge controller which handles two fixed 175W solar panels plus a portable 100W panel (so 450W max).

Question 1: I've read Will's recommendations, but I'm still confused. As part of top-balancing I pushed all the cells to between 3.630 and 3.637V. After sitting for a night they were roughly 3.52V. So which is 100%... 14.5V (max I can charge to) or 14.1V (where they eventually settle at)?

Question 2: If I want to max out my battery at ~90% SOC, what should I set the charge controller to? (Note: Renogy's default is 14.4V and it's adjustable but annoyingly only in 0.2V increments... more on this below). If I run a load and drain off 28Ah, then turn everything off and record the voltage (not sure what that is yet but let's assume for discussion purposes it's 3.4V per cell/13.6V total), is that voltage reasonably accurate for 90%? Asking since I know adding a load or a charge will vary the voltage and since I have solar panels and always at least a nominal ~0.3A load the battery will never be "idle".

Question 3: If (hypothetically) 13.6V is 90%, should I set my solar controller to 13.6V? Or should it be slightly higher to ensure it gets a full charge? It seems like the ideal charging profile would allow the solar controller to charge at 14.4V, but would cut off charging when I hit 90%/13.6V; however my Renogy doesn't support that.

Question 4: Is there a better reasonably priced charge controller I should use that is significantly more configurable? I'm not against spending ~$350 for something like the Victron MPPT 100/50, or something similar, but I only really want to spend the $ if it's going to provide a significant improvement in charging performance (meaning "speed at which it gets the battery to 90% and/or a meaningful increase in battery life). TBH I don't love my Renogy controller, but it's working and so it's hard for me to spend several hundred $ on a different for a marginal improvement... if I'm dropping the coin I have to justify it as a big improvement.

Last thing, I have a few notes about my setup:

• Because of the time involved I have not yet capacity-tested my battery pack. From what I've seen so far I expect it's at least relatively close to spec. I plan to eventually test it but it's summer and this is in my camping trailer, which I'm using for some long trips, so I don't have a two weeks available for me to drain it with my 5A load tester (measuring it along the way) and then recharge it with my 10A bench charger. So until winter when I do have the time I'm going to have to work off reasonable estimates.
• The Renogy charger boost setting is only adjustable in 0.2V increments. (charging parameters below). RIght now it's set to 14.0V, though I'm considering reducing it.
• I need to test with my multi-meter but I have about 0.15V drop between the solar controller and the battery. It's ~20' of 8 AWG wire. There's no compensation in the Renogy profile so I will have to take that into account when programming the boost voltage. Renogy's charging profile is below

 

[snoobler]

Hi folks,

New-ish guy here. I've been doing SLA and solar on my trailer for year, but I've been reading a lot of the forum posts for the last couple months and started posting recently. I finished building my LiFePO4 battery bank over the weekend, and in an effort to (over)optimize my setup I wanted to pose a couple questions to those who are smarter and far more experienced with LFP than me.

My 12V setup is 4x 271Ah CATL cells (LiitoKala "280Ah") with a pseudo-overkill solar BMS (only because overkill was out of stock). I have a Renogy Rover Elite charge controller which handles two fixed 175W solar panels plus a portable 100W panel (so 450W max).

Question 1: I've read Will's recommendations, but I'm still confused. As part of top-balancing I pushed all the cells to between 3.630 and 3.637V. After sitting for a night they were roughly 3.52V. So which is 100%... 14.5V (max I can charge to) or 14.1V (where they eventually settle at)?

Voltage rises in response to current input. When the input current is removed, the voltage falls. This varies by almost every possible parameter, i.e., each individual cell, even from the same manufacturer, can vary widely. In a few weeks, your cell voltages may vary substantially; however, they will all be between 99.5 and 100% SoC. If you happen to get cells that all settle at the same rate, that's a good sign. That's not the norm when purchasing cheap commodity cells.

LFP voltage is a poor way to determine state of charge. You need a battery monitor that counts current in and out and compares that to the capacity of the cells.

Question 2: If I want to max out my battery at ~90% SOC, what should I set the charge controller to? (Note: Renogy's default is 14.4V and it's adjustable but annoyingly only in 0.2V increments... more on this below). If I run a load and drain off 28Ah, then turn everything off and record the voltage (not sure what that is yet but let's assume for discussion purposes it's 3.4V per cell/13.6V total), is that voltage reasonably accurate for 90%? Asking since I know adding a load or a charge will vary the voltage and since I have solar panels and always at least a nominal ~0.3A load the battery will never be "idle".

This is essentially impossible to do on a voltage basis. Even at 3.4V (13.6V), you can get these to over 95% SoC. The only way to accomplish this is to establish your personal parameters based on your system, and if there's any variation on charge current (solar variability, loads, etc.), it's that much worse.

Question 3: If (hypothetically) 13.6V is 90%, should I set my solar controller to 13.6V? Or should it be slightly higher to ensure it gets a full charge? It seems like the ideal charging profile would allow the solar controller to charge at 14.4V, but would cut off charging when I hit 90%/13.6V; however my Renogy doesn't support that.

See #2. The limitation on the Renogy is common with most cheaper controllers.

Question 4: Is there a better reasonably priced charge controller I should use that is significantly more configurable? I'm not against spending ~$350 for something like the Victron MPPT 100/50, or something similar, but I only really want to spend the $ if it's going to provide a significant improvement in charging performance (meaning "speed at which it gets the battery to 90% and/or a meaningful increase in battery life). TBH I don't love my Renogy controller, but it's working and so it's hard for me to spend several hundred $ on a different for a marginal improvement... if I'm dropping the coin I have to justify it as a big improvement.

You need to be able to program a tail current (cut off). Victron absolutely has this. In using your system, you will find a given voltage and a given charge current that represents 90% on the battery monitor, and you would program your Victron to that absorption voltage and tail current. Once those values are achieved, the MPPT will cut absorption and drop to float.

This is all pretty convoluted. Another way to extend life in the same spirit is to charge slower at lower voltages. Charging to 13.8V and allowing a long slow absorption charge is less stressful to the cells and should improve cycle life in a similar manner.

 

[thegoogler]

Voltage rises in response to current input. When the input current is removed, the voltage falls. This varies by almost every possible parameter, i.e., each individual cell, even from the same manufacturer, can vary widely. In a few weeks, your cell voltages may vary substantially; however, they will all be between 99.5 and 100% SoC. If you happen to get cells that all settle at the same rate, that's a good sign. That's not the norm when purchasing cheap commodity cells.

Thanks. Two of my cells (1 and 4) seem to charge and settle at the same rate. The other 2 seem to be a few mV less. Hopefully the BMS will keep them in check.

LFP voltage is a poor way to determine state of charge. You need a battery monitor that counts current in and out and compares that to the capacity of the cells.

Yeah I do have shunts on my solar and my main 12V control center. I use the Simarine Pico for monitoring, which has coulomb counters.

One quirk is that the solar controller is near the front of my trailer (where the SLA battery was) while the shunt for the "main" (12V control/load center/converter/etc) is next to the battery which is now inside the trailer and further away. So there's a voltage drop between the solar controller and the battery, hence while the controller thinks it's putting in 10A @ 14V the battery will get a bit less due to ~20' of 8 AWG wire between them. Big question: I don't know if that affects the number of Amps delivered to the battery or just the voltage... if it's just Volts then I can compensate at the controller but if it's Amps too then my shunt will be wrong unless I move the charge controller to be closer to the batter (which then means a very long run of about 30' from the panel down to the controller.)

This is essentially impossible to do on a voltage basis. Even at 3.4V (13.6V), you can get these to over 95% SoC. The only way to accomplish this is to establish your personal parameters based on your system, and if there's any variation on charge current (solar variability, loads, etc.), it's that much worse.

Yeah I thought this is the case. Does something like Victron improve this, or don't they still use voltages to determine cutoff points?

You need to be able to program a tail current (cut off). Victron absolutely has this. In using your system, you will find a given voltage and a given charge current that represents 90% on the battery monitor, and you would program your Victron to that absorption voltage and tail current. Once those values are achieved, the MPPT will cut absorption and drop to float.

This is all pretty convoluted. Another way to extend life in the same spirit is to charge slower at lower voltages. Charging to 13.8V and allowing a long slow absorption charge is less stressful to the cells and should improve cycle life in a similar manner.

Ok I'll have to ready up more on the Victron to see how it works.

The latter is I believe what Renogy tries to do to prolong LFP battery life. Basically no float, but when the voltage drops to 13.2V kick on the charger and boost to ~14V (adjustable). Once you get there stop charging and let the battery drain again until it hits 13.2V, then start another charge cycle. And I'm thinking if I set to 13.6 or maybe 13.8V it'll be "close enough" to work. Ultimately from the CATL graphs I've seen I'm not too concerned with charging a bit over 90% (up to maybe 95%), though I have a concern that 13.2V is too low to start a new charge cycle (assuming the controller really works that way).

The chart below, BTW is from Deligreen and is supposedly the CATL 271Ah (6LH3L8). If the graph is true it seems like it's safe to run from 5-95% SOC without a significant change in voltage. The graph isn't high enough resolution but it would *appear* that at 13.6V you would only get to about 170Ah, and at 90% (243Ah) you would have to be charging somewhere around 13.8V, with 14V getting you 255-260Ah (95%).

 

[snoobler]

Thanks. Two of my cells (1 and 4) seem to charge and settle at the same rate. The other 2 seem to be a few mV less. Hopefully the BMS will keep them in check.

Again, their voltages, particularly at the XmV level are not indicative of their SoC. It may not even reflect their rate of self-discharge.

Yeah I do have shunts on my solar and my main 12V control center. I use the Simarine Pico for monitoring, which has coulomb counters.

One quirk is that the solar controller is near the front of my trailer (where the SLA battery was) while the shunt for the "main" (12V control/load center/converter/etc) is next to the battery which is now inside the trailer and further away. So there's a voltage drop between the solar controller and the battery, hence while the controller thinks it's putting in 10A @ 14V the battery will get a bit less due to ~20' of 8 AWG wire between them. Big question: I don't know if that affects the number of Amps delivered to the battery or just the voltage... if it's just Volts then I can compensate at the controller but if it's Amps too then my shunt will be wrong unless I move the charge controller to be closer to the batter (which then means a very long run of about 30' from the panel down to the controller.)

Current is not lost, but as you indicate, SCC measured voltage is higher than the shunt voltage. Another difference is that the SCC is not mentioning open circuit voltage (OCV) while the shunt is. The SCC measures the voltage as influenced by the current it's supplying. I suspect that at night, the two units read almost identically. If they don't it's a matter of accuracy/calibration.

Yeah I thought this is the case. Does something like Victron improve this, or don't they still use voltages to determine cutoff points?

Ok I'll have to ready up more on the Victron to see how it works.

Let's just say that after extensive testing, you have determined that your battery is 90% full at 14.4V and 5A of charging current. You would set the Victron to 14.4V absorption and a tail current at 5A. Once 14.4V is attained and the current is LESS than or equal to 5A for X minutes, the charger will drop to float.

The latter is I believe what Renogy tries to do to prolong LFP battery life. Basically no float, but when the voltage drops to 13.2V kick on the charger and boost to ~14V (adjustable). Once you get there stop charging and let the battery drain again until it hits 13.2V, then start another charge cycle. And I'm thinking if I set to 13.6 or maybe 13.8V it'll be "close enough" to work. Ultimately from the CATL graphs I've seen I'm not too concerned with charging a bit over 90% (up to maybe 95%), though I have a concern that 13.2V is too low to start a new charge cycle (assuming the controller really works that way).

The chart below, BTW is from Deligreen and is supposedly the CATL 271Ah (6LH3L8). If the graph is true it seems like it's safe to run from 5-95% SOC without a significant change in voltage. The graph isn't high enough resolution but it would *appear* that at 13.6V you would only get to about 170Ah, and at 90% (243Ah) you would have to be charging somewhere around 13.8V, with 14V getting you 255-260Ah (95%).

I've never met a canned LFP program that I like. IMHO, that program results in micro cycles. This unnecessarily cycles them while subjecting them to more frequent bulk/absorption charges and is likely worse for them at a float of 3.4V.

Personally, even though I love Victron, I would stick with what you have and set absorption to 13.8 and float at 13.6. Yes, you'll go to near 100% and you'll float at 95%, BUT at the higher SoC, the current will be reduced and cell stress will be minimized.

 

[thegoogler]

Personally, even though I love Victron, I would stick with what you have and set absorption to 13.8 and float at 13.6. Yes, you'll go to near 100% and you'll float at 95%, BUT at the higher SoC, the current will be reduced and cell stress will be minimized.

Annoyingly my Renogy allows me to set absorption (in 0.2V increments), but with the Lithium profile float (and equalization) is completely disabled. Boost return is fixed at 13.2V. The only way to enable float is to change the battery profile (Gel being the next closest option)

 

[snoobler]

Forgot about that. 2 hr boost is overkill for LFP and will push to 100% SoC no matter what you do.

Bin it. Go with a Victron.

 

[thegoogler]

Forgot about that. 2 hr boost is overkill for LFP and will push to 100% SoC no matter what you do.

Bin it. Go with a Victron.

I "liked" your reply but TBH I don't like the answer ;-)

So I think I need a Victron SmartSolar MPPT 100/50 plus a Smart Battery Sense sensor. Assuming I go that route anything else special I need to configure?

Any recommended resellers?

 

[snoobler]

That combo should work.

My Victron VAR is in my signature. You have to contact them for best pricing. Small outfit. They give me the best prices I can find, but they ain't no Amazon Prime.

 

[jtvt]

I've never met a canned LFP program that I like. IMHO, that program results in micro cycles. This unnecessarily cycles them while subjecting them to more frequent bulk/absorption charges and is likely worse for them at a float of 3.4V.

Annoyingly my Renogy allows me to set absorption (in 0.2V increments), but with the Lithium profile float (and equalization) is completely disabled. Boost return is fixed at 13.2V. The only way to enable float is to change the battery profile (Gel being the next closest option)

That sounds inefficient too. IIUC, the charge controller is idle until the battery is at 13.2? So even in full sun, all loads are powered by the battery? In a float mode, the CC would pick up those loads.

 

[thegoogler]

That's my understanding too. I have no idea how it'll actually work with LiFePO4. When I plugged in the battery and turned the solar controller on yesterday it immediately went into boost mode (which is normal for powering it on). I didn't hang around for 2-3 hours to wait for it to finish so I don't know how it will operate long term yet, but the manual claims there's no float mode

I had an SLA before so it would always float and in the morning typically would go through a full cycle after overnight use. I'll see if it actually works with LiFePO4 as described... if so then I probably won't be able to get a charge cycle to happen unless I turn the controller off and back on for a moment (in which case I will end up replacing it...)

Honestly the more I've used Renogy over the years the more I'm convinced their stuff is just crap

 

[jtvt]

All of them may have some weirdness since they originated out of the lead acid world. Float on Lifep04 with the Midnite classics is weird. When absorb ends, the classics won't pick up any loads until the batteries drop to the float voltage. Not a huge deal since with 56.4 absorb and 54.4 float it's usually just a few minutes. Maybe lead acids are the same, but length of time for battery voltage to drop so short that you don't notice.

 

[time2roll]

My Morningstar 60 amp MPPT is very adjustable down to 10mv with adjustable absorption time etc. Very flexible. I expect it will work fine with LFP once I get the battery in and adjust the programming.

 

[snoobler]

To my knowledge, none of the high end models have issues with LFP. They are so easily customized, you can specify suitable LFP battery parameters.

I'm not aware of others, but the Victron offers low temp charge protection to prevent charging LFP below freezing, or your desired temp.

 

[thegoogler]

To my knowledge, none of the high end models have issues with LFP. They are so easily customized, you can specify suitable LFP battery parameters.

I'm not aware of others, but the Victron offers low temp charge protection to prevent charging LFP below freezing, or your desired temp.

Thanks. LFP has proven a bit more challenging in this space than I expected, since the voltage ranges are so small. I'm sure there's a sales (integration) opportunity here for some manufacturer to integrate communications between a battery SoC monitor, solar charge controller, and LFP BMS.

Personally low temp cutoff isn't important to me since I have a BMS that will handle that (and in fact I might prefer the solar controller still operates below freezing because if there's sunlight it can still help partially power my normal 12V loads until the battery warms up), though I know other people in the world exist and might be reading this.

 

[Daddy Tanuki]

yeah I use Morningstar for my solar controllers... but they are rather iffy on answering questions about how to set them up... they keep falling back on several useless reply's "oh we don't do that, oh we are not sure, oh maybe you should get a different battery system, inverter, BMS etc. never any actual answers, just deflection and a ask somebody else type of attitude. the units can do it according to the manuals... but tech help... yeahright.

 

[blutow]

I haven't dug into all the advanced functionality (assistants, etc.) of my victron equipment, but it seems like it's all voltage based to trigger charge cycles and that just seems fundamentally flawed to me for LFP.

I get that you may not be able to count exclusively on a shunt/capacity monitor to manage charging, but it just seems like it should be part of the trigger. From the testing I've done, I can run a really small load on the system for hours and pull significant capacity out and never trigger charging. Or, I can run a really large load for a short period and only pull 1% of the capacity out of the system, but that will trigger a charge cycle because the voltage dropped a bit with the large load. I don't know if those exact numbers are right, but that's the general behavior I see. Charging sometimes when it doesn't need to charge and not charging when it should. It seems that the victron SOC calculation is very good at the top end (constantly resetting when the system is fully charged). Why not trigger a charge cycle any time the SOC drops below x% (or voltage drops below y) rather than trying to do it strictly off system voltage without consideration for SOC? You can continue using voltage to manage the actual charge cycle, but voltage seems to be a poor trigger to start it. Maybe this is already possible, but I have not found it.

I'm no expert on this stuff, but everything I read and see with my system is that trying to determine SOC from voltage is a fools errand, but all the chargers seem to trigger charge cycles off voltage. It just seems like charging could be smarter and consider SOC as one of it's inputs.

 

[Substrate]

yeah I use Morningstar for my solar controllers... but they are rather iffy on answering questions about how to set them up... they keep falling back on several useless reply's "oh we don't do that, oh we are not sure, oh maybe you should get a different battery system, inverter, BMS etc. never any actual answers, just deflection and a ask somebody else type of attitude. the units can do it according to the manuals... but tech help... yeahright.

I've run into that before, but I'm pretty sure I know why. Nearly all SCC's out of the box are on the safe-conservative side, and as such are programmed for a standby type of application. Protects the company and the unskilled consumer, which tend to start out with older batteries that have seen better days. At least mostly from the lead-acid environment.

Daily-cyclic however is another issue, and they simply can't provide the engineering support, nor corporate protection from the person who goes totally bananas, and fries his battery. Nor bicker back and forth about what may come from a forum or other entity when that happens.

So not a bad thing - I'm pretty sure most companies simply don't have the resources to engineer all the customer's systems if they want to deviate from the OOB settings. I'd hate to be on the other end of that phone/email train all day.

 

[time2roll]

Yes the parameters of Morningstar are exceptional for lead-acid especially as it relates to automatically adjusting to storage vs in use.

There is no charge it up and turn off for LFP. Or I want a low storage float and absorption voltage unless it senses additional use and then charge to a higher voltage.

Some day.